IT India Cornell Rev

7/28/2019 IT India Cornell Rev

1/32


2/32

Information Technology and Indias Economic Development Nirvikar Singh, UC Santa Cruz, July 2002

Information Technology and Indias Economic Development

Nirvikar Singh

Professor of EconomicsUniversity of California, Santa CruzSanta Cruz, CA 95064, USA

Revised, July 2002

Abstract

This paper discusses the possibilities for broad-based IT-led economic growth in India, includingincreasing value-added, using better telecom links to capture more benefits domestically through

offshore development for developed country firms, greater spillovers to the local economy,broadening the IT industry with production of telecom access devices, improving the functioningof the economy through a more extensive and denser communications network, and improvinggovernance. We also examine the policy environment, arguing that government policy is betterfocused on removing labor market distortions and infrastructure constraints, rather thanproviding output or export subsidies to the software industry.

Keywords: information technology, software, complementarities, telecommunicationsJEL Classification: M21, L63, O12, O3

This paper is a revision of one presented at a conference on the Indian Economy, held at Cornell on April 19-20,2002. I am grateful to Kaushik Basu, Yale Braunstein, Ashok Desai, Atanu Dey, Rafiq Dossani, Kyle Eischen, P.D.Kaushik and Vini Mahajan for helpful discussions and comments on related work. I am particularly grateful toDevesh Kapur for his comments as a discussant at the conference. None of them is responsible for any remainingshortcomings. I have learned about aspects of this topic from numerous people, in addition to the ones named. I havetried to reference their works as much as possible, and regret any inadvertent omissions.

1


3/32


1. Introduction

The success of Indias software industry on the global stage has captured the imaginationof Indians in a way that only cricket and hockey successes could in the past. Indians (or peopleof Indian origin) have become leaders of, as well as contributors to, the information technology

(IT) revolution in the United States, reinforcing the impression that India is world class in IT. Atthe same time, India remains an extremely poor country, with levels of human development forthe masses that put it in the same league as sub-Saharan Africa. From this perspective, Indias ITsuccess represents the emergence of another elite enclave, with increased inequality the result.

In this paper, we examine the question of whether IT can do more than fuel an enclave-based export boom. Can IT contribute to Indias economic development in a broader, morefundamental way? What are the potential mechanisms whereby this can occur? What is thelikelihood of IT accelerating Indias growth, and what are the potential roadblocks or bottleneckswhere government policy can make a difference between success and failure? This paperassumes a basic familiarity with the general structure and performance of the Indian economy,

and the economic reform process that has been taking place through the last decade or more. InSection 2, therefore, we begin our analysis directly by examining the performance of Indias ITsector, discussing the role of software versus hardware, the growth pattern of the softwareindustry and software exports, the rapid emergence of IT-enabled services, and the role of thedomestic market.

Section 3 turns to a consideration of the resource needs of the IT sector, and possibleconstraints and bottlenecks. These include the supply of IT-skilled labor to support futuregrowth, telecommunications and other aspects of infrastructure, and possible financialconstraints. Section 4 uses a range of economic ideas to map out the possibilities for broad-basedIT-led development, going beyond the IT sector. We draw on recent analyses of the process of

economic development that emphasize factors such as innovation, complementarities intechnologies and in demand, and pecuniary externalities. In terms of the mechanisms fordevelopment, we discuss examples such as increasing value-added, using better telecom links tocapture more benefits domestically through offshore development for industrial country firms,greater spillovers to the local economy, broadening the IT industry with production of telecomaccess devices, improving the functioning of the economy through a more extensive and densercommunications network, and improving governance at all levels.

Section 5 examines the policy environment, which interacts with resource availability, inthe light of broader developmental possibilities. Issues raised here include the provision ofeducation, labor market distortions, infrastructure development in areas such as

telecommunications, and tax and subsidy policies. Section 6 provides a summary conclusion,with an assessment of possibilities and recommendations for policy.

2


4/32


2. The IT Sector

Information technology essentially refers to the digital processing, storage andcommunication of information of all kinds1. Therefore, IT can potentially be used in every sectorof the economy.2 The true impact of IT on growth and productivity continues to be a matter ofdebate, even in the United States, which has been the leader and largest adopter of IT.

3

However, there is no doubt that the IT sector has been a dynamic one in many developedcountries, and India has stood out as a developing country where IT, in the guise of softwareexports, has grown dramatically, despite the countrys relatively low level of income anddevelopment. An example of ITs broader impact comes from the case of so-called IT-enabledservices, a broad category covering many different kinds of data processing and voiceinteractions that use some IT infrastructure as inputs, but do not necessarily involve theproduction of IT outputs. Indias figures for the size of the IT sector typically include suchservices, and they will be discussed in this section. We begin with a review of the overallindustry size, then discuss software versus hardware, exports versus domestic sales, and, finally,IT-enabled services.

Table 1: Indias GDP and IT Sector

Year

GDP at

current prices

(Rs. Billion)

IT sector

(Rs. Billion)

IT sector

(US $ Billion)

1994-95 9,170 63 2.0

1995-96 10,732 99 2.9

1996-97 12,435 137 3.8

1997-98 13,900 187 5.0

1998-99 16,160 248 6.1

1999-00 17,865 371 8.7

2000-01 19,895 554 12.2

Sources: GDP: www.adb.org, IT sector: www.nasscom.org/it_industry/indic_statistics.asp

1 A popular alternative is ICT, for information and communications technology: the World Bank, for example,favors this term.2 In this sense, it is an exemplar of what is called a general-purpose technology in economic modeling of growthprocesses. See Helpman (1998) and Kapur and Ramamurti (2001). We take up this idea in detail in section 4.3 To give a sampling of research in the US, David (2000) emphasizes the lag with which any new technology affectsproductivity; Gordon (2000) offers a skeptical view of the impact of IT on productivity, arguing that the empiricalevidence indicates that the impact is narrow and limited; Jorgenson, in the most comprehensive analysis, finds thatIT has contributed significantly to total factor productivity growth (TFPG) in the US. Of course, higher TFPGimplies higher overall growth, ceteris paribus.

3


5/32


Table 1 provides statistics on the overall size of GDP, and on the size of the IT industry,in billions of current Rupees. The IT industry figures are not necessarily as accurate as the GDPfigures, and they are based on revenues rather than value added, so they are not conceptuallydirectly comparable. Nevertheless, Table 1 gives an idea of the growing importance of the ITsector in Indias economy. The IT sector grew over this six-year period by a factor of nine,

whereas GDP slightly more than doubled in the same period. Table 1 also gives dollarequivalents for the IT sector figures: these reflect the changing exchange rate over the period,though again there appear to be some discrepancies. With the Rupee falling against the US dollarover this period, the dollar value of the IT sector grew by a factor of six.

Even with its spectacular growth, Indias IT sector in 2000-01 was less than 3% of GDP(even if one treats all sales as value added), suggesting that there is considerable room for furthergrowth. For example, if Indias nominal GDP were to double over the six years subsequent to2000-01, while the IT sector were to increase by a factor of six, and if we assume that valueadded is about two thirds of total sales, the resulting ratio of the IT sector to GDP would be inthe range of 8%, or not dissimilar to that in the US today. This calculation assumes slower

growth than in the 1999 McKinsey-NASSCOM projections, but may be a more likely figure.

Software vs. HardwareThe basic distinction in IT is between hardware and software. The former refers, of

course, to the physical components of processors, storage devices and communications devices.The latter refers to the instructions that govern the flow and processing of information in digitalform, within and between hardware devices and components. The design of hardware actuallyinvolves the development and use of appropriate software code, so there are definite overlaps inthe two categories. It is also possible to substitute software for hardware in the basic design ofcircuits. The actual production of hardware is classified within the manufacturing sector, and ismore distinct from the development of software. Profitably manufacturing semiconductors and

other sophisticated hardware components typically requires infrastructure, large-scaleinvestments in capacity, and accumulated experience that India does not possess, and is not in aposition to acquire easily. Indias development path, despite its emphasis on import-substitutingindustrialization, has not supported the growth of a robust, world-class manufacturing industry,such as has arisen in many East Asian countries.

Nevertheless, India does perform many hardware assembly tasks internally, almostentirely for the domestic market. Components in such cases typically come from East andSoutheast Asia. The ability to organize this aspect of production may be the basis for furtherdevelopment of hardware capabilities. Several East Asian countries also began as mainlyassemblers of sophisticated components produced elsewhere, and extended their presence in thevalue chain backward as they learned by doing. While being late to the game may make entrymore difficult, the fact that manufacturing of components becomes increasingly standardized,and the cost of these components falls, works in favor of late entrants. For example, theproduction of most memory chips has become commoditized, and moved to developingcountries, where 20 years ago it was the core of Intels business.

The example of firms like Dell and Cisco may also be useful to keep in mind whenevaluating the hardware industry in India. Dell outsources most, if not all, of its component

4


6/32


manufacturing. It is, in fact, an extremely sophisticated assembler. Its value creation is based onorganizing this assembly as efficiently as possible, doing so on demand, and keeping itsinventories absolutely minimal. Strong customer service plus management of communicationsand logistics at both ends of the value chain are also keys to Dells success. Dells positioning totake advantage of strengths in infrastructure and closeness to a growing customer market is an

important lesson for India. The point here, which we shall further develop subsequently in thispaper, is that a hardware industry in India is feasible, building on Indias experience in assembly,but it will require significant improvements in infrastructure, and careful attention to marketneeds. Our earlier discussion of hardware design is relevant here, since this is an area wheresoftware expertise matters more, and manufacturing infrastructure does not. Therefore hardwaredesign may be a promising area for Indian IT. Table 2 gives some basic statistics on thedecomposition of Indias IT sector.

Table 2: Indias IT Sector Decomposition (Rs. Billion)

Year Hardware Software Other* Domestic** Export**

1994-95 23.8 26.1 13.6 34.6 21.1

1995-96 36.8 41.9 20.2 59.0 26.6

1996-97 48.1 63.1 25.8 68.4 49.8

1997-98 52.4 100.4 33.8 88.5 73.4

1998-99 42.5 158.9 36.4 105.4 110.3

1999-00 65.7 243.5 61.6 152.7 176.3

Sources: www.nasscom.org/it_industry/indic_statistics.aspNotes: *Includes peripherals, training, maintenance and networking; **Hardware, software, peripherals.

Indias software industry is more robust than its hardware industry, at least in certainareas. While selling packaged software to consumer (and most business) markets requireseconomies of scale and scope, as well as marketing and customer support muscle, project-oriented components of software development do not do so, to quite the same degree. Thesoftware development and use life cycle includes analysis and specification of requirements,design, coding, testing, installation, maintenance and support. Many of these activities,particularly coding and testing, involve relatively routine IT skills that Indias workforce has in

large absolute numbers (though small relative to the total population).

The existence of the Indian Institutes of Technology (IITs), the ubiquity of Unix inacademic environments, and the relatively low infrastructure demands of learning to use andcreate software all worked in Indias favor on the supply side. The use of English in Indiashigher education system, the increase in the use of Unix and related operating systems due to theexplosion of the Internet, and the large number of Y2K-related projects in the late 1990s allcontributed to demand for Indias software industry services, in addition to the general growth in

5


7/32


IT in the 1990s. Much of this demand came from abroad, as we discuss in the next subsection.However, the software industrys domestic revenues also grew rapidly in the last few years, atover 30% per annum, on average.

Despite the even faster growth of software exports, domestic software revenue still

represents about 30% of software industry gross receipts. The National Association of Softwareand Services Companies (NASSCOM) projects domestic sales to grow substantially faster thanexport sales in the next decade, enough to make domestic sales over 50% of the industrys sales,and this seems to borne out by recent growth in domestic software (31% annual growth from1998-99 to 2000-01, NASSCOM, 2002a, b), and IT overall (42% growth in 1999-00). Much ofthis growth was driven by demand from state governments, which have been introducing e-governance initiatives. The second largest sector in terms of domestic IT demand is financialservices. In the domestic market, products and packages make up almost half of revenues, withprojects accounting for over a quarter, while professional services, support and maintenance, IT-enabled services and training each make up less than 10% (Singh 2002). This pattern is quitedifferent from that of exports, as we discuss below.

Software ExportsIndias software exports went from a few million dollars in the 1980s, to over a billion

dollars by 1995-96, and $6.2 billion in 2000-01. Despite the global slowdown, software andservices exports exceeded the 2000-01 figure by 31% in dollar terms, crossing $8 billion.4 Whilegrowth has slowed from the earlier 50% rate, the dollar increases are as big. While theNASSCOM-McKinsey target of $50 billion in software export revenue by 2008 may beoveroptimistic,5 the resilience of software exports in a difficult economic climate has provedpessimists wrong.

The pattern of activities that generates software export revenue is somewhat different

from the sources of domestic revenue. In particular, professional services accounted for 44% ofexport revenue in 1998-99, followed by projects at 36%. Products and packages were only 8%of the total compared to nearly half for domestic revenue. Desai (2000) suggests that thedifference in the patterns of domestic and export sales is overstated, because domestic sales ofpackages are by resellers of packaged software licensed from foreign software vendors.However, this appears to be changing, as Indian firms develop packages for the home market inareas such as financial services.

Coding and testing appear to form a significant proportion of the skills used in the Indiansoftware industry. Some have expressed concern6 that the Indian software industry isprogrammer heavy, and therefore unable to move up to higher value-added segments ofsoftware. Related issues that reinforce these concerns are the brain drain of the most talented orexperienced IT people, the lack of sufficient managerial skills for more sophisticated contractwork, and the lack of domestic spillovers from the body shopping of programmers for onsite

4 Data from Electronics and Computer Software Export Promotion Council, reported in Business Standard, April 9,2002. Hardware export figures, while much smaller, also showed robust growth.5 See http://www.nasscom.org/it_industry/sw_export.asp.6 See, for example, Heeks (1996, 1998) and Desai (2000).

6


8/32


work in developed countries. Again, the latest trends suggest that Indian firms have begun toovercome these roadblocks.

7

Indicators of the strength of Indias software export capabilities include the depth of itsbase, and the breadth of its global reach. There are over 2,000 Indian software exporters, and

while only the top five (TCS, Infosys, Wipro, Satyam and HCL) are or are approaching thestatus of global brands, they together account for only about 30% of software exports.8 Whilethe United States remains by far the largest market for Indias software exports, its share ofIndias software exports has fallen slightly, to 62%, with Europe coming in at 24%, and Japanand the rest of the world accounting for the remaining 14%.9 Individual firms and organizationssuch as NASSCOM have shown themselves to be adept at targeting markets with substantialgrowth potential, such as Germany, and the reputations built in exporting to the US are provingimportant.10

IT-Enabled ServicesIT-enabled services are not necessarily related to the production of software or IT in

general, but use IT to make the provision of services possible. Customer call centers are oneexample, where Indians have been training to speak with American accents, in order to deal withcustomer queries from the US. Accounting services are a second example. Yet another, morelong-standing market segment is that of medical transcription. NASSCOM provides acategorization of 10 different types of IT-enabled services (Table 3), varying widely in terms ofskills required and value added.11 The ten categories overlap to some extent, but they give a goodidea of the scope of the industry.

Table 3: IT-Enabled Services Types

Customer Interaction Services

Business Process Outsourcing /Management; Back Office Operations Insurance Claims Processing Medical Transcription Legal Databases Digital Content Online Education Data Digitization / GIS Payroll / HR Services Web site Services

7 See, for example, Tschang (2001) and Kapur (2002).8 These figures are taken or constructed from http://www.nasscom.org/it_industry/top20_exporters.asp. See alsohttp://www.nasscom.org/it_industry/top20_sw_cos.asp and http://www.nasscom.org/it_industry/sw_export.asp.9 The figures are from http://www.nasscom.org/it_industry/export_destinations.asp Similar figures are given by

Dataquest(2001), Heeks (1998) and Desai (2000).10 For example, see http://www.tcs.com/news/tcs_media/htdocs/sh01/nov01_FT_article.htm11 Raman Roy, CEO of Spectramind, suggests five categories of teleworking: data entry and conversion, rule-setprocessing, problem-solving, direct customer interaction and expert knowledge services, ranked in terms ofincreasing sophistication and value added. See The Economist(2001), p. 60.

7


9/32


Source: http://www.nasscom.org/it_industry/spectrum.asp

In terms of potential, the first category in Table 3 is projected to be reach $60 billionworldwide in 2003. Other categories may be similar in size or smaller, but Indian industry ismaking inroads into all of them. According to NASSCOM, revenues from IT-enabled services inIndia in 2000-01 were over $ 800 million, up 70% from the previous year. Customer interaction

services were the highest growing segment within this sector with over 100% growth in 2000-01,and generating 20% of the sectors revenue. Back office operations revenues grew over 40% inthe same period, and provided a third of the sectors revenues. NASSCOM projects IT enabledservices to generate revenues of $ 17 billion and provide employment for 1.1 million people by2010, while a report submitted to the Indian governments Electronics and Computer SoftwareExport Promotion Council, sees IT-enabled services exports growing from their estimate of $264million in 2000 to over $4 billion in 2005.12.

The rapid growth of the sector illustrates that many of the early problems have beenovercome. These included firms that were too small to market to or deal directly with clients,and to make investments in adequate training and infrastructure.13 Success has generated

attention, and NASSCOM, the government and others are working on providing incentives,venture funding, training and infrastructure. Good communications links are obviously importantfor the success of IT-enabled services. An additional bottleneck in the past may have been thelack of managerial and marketing skills, and of reputations for quality. Part of the solutionincludes the import of such skills by multinationals such as GE and Citigroup shifting some oftheir back office operations to India. There is no doubt that there have been positive reputationaland resource spillovers from the software export sector to IT-enabled services.14

3. Resource Considerations

As in any industry, the availability of adequate supplies of inputs is critical for growth.Much of the caution about the prospects for Indias IT industry has been focused on potentialbottlenecks in the supply of skills, and the quality of the infrastructure. We add financialconstraints to this combination, and discuss each of these briefly.

Supply of SkillsA major reason for the success of Indias software industry is the large supply of labor

with some IT skills. India graduates perhaps about 125,000 engineers a year, second only to theUS worldwide.15 However, not all these engineers go into the IT industry, and not all ITprofessionals have engineering or computer science qualifications this being true of the US as

12 See The Economist(2001).13 SeeDQ Week(2001).14 See Kapur and Ramamurti (2001) for a more detailed argument, including the role of Indians with experience ofworking in Silicon Valley and other global IT centers.15 SeeBusiness Week(2001). However, Aggarwal (2001) gives a substantially lower figure of 55,000 engineeringgraduates annually, excluding private institutes and Masters of Computer Applications (MCA). Arora andArunachalam (2000) estimate an overall figure, including MCAs and graduates of informal training institutes, ofclose to 140,000 per annum. Kapur (2002) quotes World Bank estimates of 160,000 graduates and diploma holdersin engineering and technology in the late 1990s. See also Arora et al (2001a, 2001b) and Saxenian (2001).

8


10/32


well.16 Indias stock of IT professionals is estimated at over 400,000,17 so that IT industryrevenues per IT professional (assuming that all of them work in the industry, which is unlikely)are about $30,000.18 Government targets and others optimistic projections imply IT industryrevenues will increase by a factor of 15. The breakdown of this growth could be something likea doubling in revenue per IT professional, and therefore almost an eightfold increase in numbers.

However, to the extent that much of this future growth will come in IT-enabled services, theadditional employment may come in areas where different, easier-to-acquire skills are needed.

Whether growth comes in revenue per employee or number of workers, there areimplications for training. Increasing revenue per IT professional requires improvements inmanagerial and marketing skills19, as well as the production of more highly trained IT people.Training more people in IT requires investments to increase the capacity of this component ofthe higher education sector. This is a thorny problem, given the poor state of Indias highereducation system. Even at the elite IITs, faculty are poorly paid relative to industry, and thephysical infrastructure has deteriorated from lack of investment. Increases in governmentexpenditure to fix these problems are difficult in an environment of large budget deficits and

long-term neglect of basic education. Interestingly, the task of training workers for IT-enabledservices such as call centers is much easier, since the universities produce large numbers ofgraduates with some familiarity with the English language and Western culture. In all areas of ITand related services, however, increased private and public investment has occurred.20 Onepotential problem is that of maintaining standards and quality, but industry-determinedcertifications exist, and reputations can be established quite quickly in practice in a competitiveenvironment. IT industry investment itself plays a role, since the industry has a strong interest ingrowing the available supply of IT professionals.21

A further problem besides sheer numbers is the issue of level of training, and even theIITs are hard pressed to provide postgraduate education comparable to what is available in theUS. Here, the brain drain, initially severe, can be beneficial, as the current slowdown in the USsends back thousands of Indian IT professionals with valuable training and experience.22 Desai(2000) uses this issue to suggest that India may actually be better off by continuing to specializein the lower end of the market, for coding and testing, as well as in IT-enabled services, at leastin the next few years, but this argument may miss the point that sustained success requires thesimultaneous use of workers with many complementary skills at different levels.

23Which route

16 See Arora et al(2000), as well as Heeks (1996) and Desai (2000) for further discussion.17 Data from NASSCOM, reported in CCI Business Bulletin, February 23-March 1, 2002,http://www.ccindia.com/bulletin.html.18 This uses the revenue figure from Table 1, for the Indian IT industry, and is overstated to the extent that itexcludes some types of employees. Arora et al(2001a) construct a lower estimate of $15,600 for 1998-99 (their

Table 1).19 The implication is that changes in the product-service mix toward that involving higher value-added tasks wouldbe associated with these improvements, resulting in increased productivity.20 See Arora and Athreye (2002) and Kapur (2002) and the references therein. The former paper emphasizes theregional concentration of engineering colleges in India.21 IT industry figures include over $400 million of IT training revenues, for example.22 Kapur (2002) argues that this process has begun and been important in India, though perhaps not to the sameextent as in countries like Taiwan (Saxenian, 2000).23 This idea has been formalized in the O-ring theory of production, developed by Kremer (1993). See also Basu(1997), Chapter 2.

9


11/32


12/32


international bandwidth constraints.26 Several problem areas remain, which will requireattention. These include the system of interconnect charges, licensing fees and depositrequirements for entry, restrictions on franchising, bandwidth allocations, and so on. Thechallenge of building a financially viable, robust and extensive telecoms infrastructure stillexists.

27The tendency of government regulators still appears to be to over regulate, one

prominent example being the requirement for new private telecoms to meet old-style quotas forinstalling village telephones, without adequate regard to financial viability.

With appropriate policy adjustments, technological progress, including domesticinnovation, may be an important factor in removing current telecoms infrastructure constraints.Ashok Jhunjhunwala (footnote 27) gives the example of cable services in India, which are pricedat $2 to $4 per month, and have 35-40 million subscribers. At this kind of price point, however,a rural telecoms operator in India cannot recover set-up costs for access, which are about $800using conventional technologies. The goal of innovations by Jhunjhunwalas team, therefore, hasbeen to bring the cost of combined Internet and voice access down to $200. The latter figurewould make access affordable to 50% of Indian households at current income levels. Without

such innovations, government targets of increasing Indias teledensity fourfold (from 4 to 15 perhundred), or Internet access tenfold are empty rhetoric.

The bottom line is that bringing down the cost of access through innovation targeted atthe domestic market is a critical component of any dramatic increase in telecoms connectivity inIndia. Economically combining Internet and voice access also has the benefit of increasing thevalue of connecting to the network. The benefits accrue not just to the poor, but also to the tensof millions of lower middle class households who are currently outside the affordability radius.A denser domestic network not only increases the value of international network links, but it alsoprovides opportunities for increasing the rate of training of IT personnel. Finally, thedevelopment of an indigenous hardware industry for low-cost access devices and networkcomponents has the potential to fill in gaps in Indias IT capabilities on the manufacturingfront.28

Financial Constraints

A striking feature of the Indian economy pre-reform was its inefficient use of capital.Relatively high savings rates were associated with relatively low growth rates. Financial sectorreform in India has focused on making the countrys organized capital markets more efficient.Simple institutional improvements such as electronic trading and settlement, guidelines forcorporate governance, and so on, have been introduced. However, the nature of the financialsystem overall still involves financial repression, with the banking sector and a large number ofother financial institutions being subject to parking of government and state enterprise deficits

and to directed lending.

29

These problems mean that substantial inefficiencies remain in the

26 See Singh (2002) for further details and additional references.27 Two sources for tracking policy issues and broader concerns are www.trai.gov.in, the website of the TelecomsRegulatory Authority of India, and http://www.tenet.res.in/Papers/papers.html, which features the work of the IIT-Chennai group headed by Ashok Jhunjhunwala. The most recent of these is www.tenet.res.in/Papers/techolo.html.See also Jhunjhunwala (2000).28This does not mean trying to do it all in-house: again Dell is a useful example to bear in mind.29 See Singh and Srinivasan (2002) for a more detailed discussion in the context of federalism and reform.

11


13/32


financial system. This has negative implications for industry overall30, but particularly for a fast-growing sector such as IT.

Clearly, broader reform of the financial sector is required. While such reform has, asnoted, been taking place in areas such as the functioning of Indian stock markets, corporate

governance, regulation of banking, and methods of central government borrowing, theconstraints imposed by the web of government-controlled financial institutions and their badloans to the public sector are a severe hurdle to further reform. For IT start-ups, venture capitalhas been extremely important, and this should be the case for India also. While the initial lack ofa venture capital industry in India may have been positive, in the sense that the policies to createone could be considered from scratch, efforts to do so have tangled with existing mazes offinancial regulations and legal restrictions, including tax and corporate law. An importantbeginning was made by a committee on venture capital appointed by the Securities andExchange Board of India (SEBI), Indias chief financial regulator. The committees report wasadopted by SEBI in June 2000, but many of the changes required are beyond SEBIsjurisdiction.31

Despite policy hurdles, which are still receiving attention, venture capital in India isstarting to take off. A government sponsored VC fund, the National Venture Fund for Softwareand IT industry (NFSIT), was launched in December 1999. States such as Andhra Pradesh,Karnataka, Delhi, Kerala, Gujarat, and Tamil Nadu have also set up their own venture funds. It isnot clear how effective government-sponsored funds can be, since venture capital involves highrisks that are not normally associated with government activities, and government interventionmay be subject to other incentive problems. Putting aside these issues, it is true that venturecapital funding in Indias IT sector increased from $80 million in 1997-98 to $500 million justtwo years later.32 If a venture capital industry can flourish, and stock market institutions cancontinue to develop, the growth of Indias IT sector can be fueled, but the problems of the rest ofthe financial sector still cast a shadow.

4. IT and Development

The case for IT as an engine of growth and development must mainly rest on standardeconomic criteria, such as comparative advantage, complementarities, and the dynamics of theglobal economy. The IT sector can be an important source of growth for India if the country hasa comparative advantage in providing certain kinds of IT-related products and services, if theglobal demand for these products and services is likely to grow rapidly, and if the growth of thesector has positive spillover benefits to the rest of the domestic economy. The first two of theseconditions seem to be well established, though they merit some discussion of future possibilities,

particularly with respect to the reasons for and the dynamics of Indias comparative advantage inthis sector. One of the most interesting issues, which we wish to emphasize here, is the third

30 It is arguable that the problem of low growth in Indias manufacturing is substantially attributable to difficulties infinancing investment.31 Important overviews of the issues are in Dossani and Saez (2000) and Dossani and Kenney (2001). Rafiq Dossaniwas one of the members of the SEBI committee.32 The data, and other information on Indias VC environment, can be found athttp://www.nasscom.org/business_in_india/vc_scenario.asp.

12


14/32


condition, of spillover benefits. This is the area where the IT sector may be special, and not justanother export enclave. Furthermore, IT may have a role to play in broader human development,beyond just economic growth. This is a contentious issue, with sharply opposing viewsexpressed. We will proceed in this section as follows. First, we outline some theoretical ideasthat are relevant for thinking about the role of IT in growth, and which will inform our

consideration of different aspects of this role. Thus, we sequentially examine issues ofcomparative advantage in software and services, the development of a domestic market,spillovers to the economy as a whole, and the potential impact on governance.

IT and Growth Theories

The starting point for considering the role of IT in development has to be theories ofgrowth that give endogenous innovation a central role. The ingredients of these models typicallyinclude differentiated capital inputs, monopolistic competition, production of new inputs throughR&D, and ultimately economy-wide increasing returns that allow sustained growth to occur.Hence these models shift away from the exclusive focus on capital accumulation thatcharacterized the neoclassical growth model and the core of Indian post-independence economic

policy. The work of Grossman and Helpman (1991) and Rivera-Batiz and Romer (1991a,b)incorporates international trade and the evolution of comparative advantage into endogenousgrowth models. In these analyses, the economy is typically divided into manufacturing, R&Dand traditional sectors, so IT does not necessarily fit neatly into any one category. Design anddevelopment of software may have characteristics of R&D, while IT-enabled services are morelike manufacturing in their use of established techniques for production. The general message ofthese models, however, is that externalities associated with monopolistic competition may givepolicy a role in influencing the evolution of comparative advantage.33

The general models of endogenous growth leave open the issue of what makes IT special.Here, the concept of general-purpose technologies (GPTs) seems very useful. The idea of GPTs

was introduced by Bresnahan and Trajtenberg (1995), who define them in terms of having threekey characteristics: pervasiveness, technological dynamism and innovationalcomplementarities.34 Examples of GPTs include writing and printing (both earlier advances inIT), modern digital electronic IT, steam and electricity (both advances in power deliverysystems) and synthetic materials. Helpman and Trajtenberg (1998a, 1998b) have developed amodel of growth led by GPTs, in which sustained growth comes from the periodic, exogenousintroduction of new GPTs. Other mechanisms that would give endogenous growth are ruled out,but otherwise, the framework consisting of endogenous R&D, monopolistic competition and theintroduction of new intermediate inputs as the mechanism of growth is similar to endogenousgrowth models. In these models, any GPT has similar abstract effects. Dudley (1999), using adifferent theoretical approach, but the same overall idea, makes a case for information and

communication technology as a particularly influential or fundamental GPT. He constructs asimple model of innovation working through falling costs of communication in networks.

33 Other, more recent treatments of endogenous growth and trade that may be useful in thinking about IT in Indiainclude Basu and Weil (1998) and Chuang (1998).34 See Lipsey, et al(1998) for a detailed survey and examination of the concept, as well as the other pieces inHelpman (1998). The idea of complementarities is discussed in greater detail below: it is closely related to the olderidea of linkages: see Basu (1997) and Ray (1998) for references and discussion.

13


15/32


We can say a little more about the characteristics of GPTs in the context of IT inparticular. Pervasiveness seems to be potentially a natural property of IT. In the Indian context,doubts are centered on issues of cost and access. We have touched on those briefly earlier in thepaper, and will return to them later in this section. Table 4, however, illustrates the importantpositive trends that support pervasiveness. Technological dynamism refers to the potential for

sustained innovation that come with new GPTs, and is again illustrated by the dramatic fall incosts shown in Table 4. Complementarities of GPTs are vertical complementarities, becauseGPTs spur innovation and lower manufacturing costs in downstream sectors, with positivefeedback effects to the GPT itself. There are also horizontal complementarities, since thedownstream sectors may face a coordination problem in expanding sufficiently to encourage theimprovement of the GPT. Note that international trade with a more advanced country may be oneway to overcome some of these externality problems. That seems to the lesson of Indias ITsector development.

Table 4: Falling Costs of Computing ($)

Costs of computing 1970 1999

1 Mhz of processing power 7,601 0.17

1 megabit of storage 5,257 0.17

1 trillion bits sent 150,000 0.12

Source: Pam Woodall, The New Economy: Survey, The Economist, September 23, 2000, p. 6, Chart 1.

The importance of complementarities in understanding growth processes has been

described in most detail by Matsuyama (1995, see also Ciccone and Matsuyama, 1996, Basu,1997, and the references therein). Matsuyama makes three useful observations. The first is theidentification of the differing roles played by horizontal and vertical complementarities, such aswe discussed in the previous paragraph. The second is the difference between technologicalcomplementarities, emphasized by writers such as Kremer (1993) and Milgrom, Qian andRoberts (1991) and the demand-based complementarities and pecuniary externalities that drivemodels such as those of Matsuyama and his co-authors. The third point is the difference betweenthe effects of history and of expectations in affecting equilibrium outcomes and growth. Either orboth may work against development and growth, by preventing coordinated movement out of abad equilibrium.

Matsuyama examines a range of models, and shows how growth may be arrested orsustained, and what kinds of inefficiencies might arise. In particular, the externalities generatedby the structure of complementarities can lead to inefficiencies that are best characterized ascoordination failures. This set of problems also arises in the GPT models of Helpman andTrajtenberg, discussed above. Without going into details, we suggest that this literature has somerelevance for thinking about the role of IT in Indian development. In particular, while the successof IT so far may be the result of factors that have to do with initial comparative advantage (seebelow), the fortuitousness of freedom from government controls (see Kapur, 2002) and

14


16/32


integration with the world economy during the boom of the 1990s, the kinds of problems that ITmay face in the future, as an engine of growth, have to do with potential coordination failures inproviding other inputs along with IT, or in the downstream sectors that use IT. For example, ifIndian manufacturing remains moribund because of the governments fiscal problems and theireffects on the financial sector, a significant market for Indias IT sector may be stifled. We

investigate some of these issues in the remainder of this section.

Comparative AdvantageThe static theory of international trade is based on comparative advantage, determined by

relative factor endowments and/or technology differences. Empirically testing this theory isdifficult, since other influences may also be at work. For example, intra-industry trade driven byproduct differentiation and economies of scale may involve different trade patterns than thosebased on traditional comparative advantage. However, in the case of software exports,attributing the export boom to comparative advantage seems reasonable.35 We have noted Indiaspool of workers with software and language skills that are valued in the international market:they are the source of Indias comparative advantage in at least some segments of the software

industry. As Kapur (2002) emphasizes, the lack of explicit government restrictions on this sectorallowed this comparative advantage to assert itself, whereas manufacturing has not enjoyed thesame benefits.

While India missed the boat with respect to the labor-intensive manufactured exports thatcontributed to the East Asian miracle, it is now in a position to replicate this phenomenon withlabor-intensive software services and (even more labor-intensive) IT-enabled services. Even ifexports of this nature cannot sustain earlier growth rates, they can make a substantialcontribution. For example, 20% growth in a sector that is 5% of the economy adds onepercentage point to overall economic growth. In the very short run, therefore, moving up theladder of value added (or establishing a broader hold on the value chain) may not be a critical

issue.

There are two reasons for not stopping here, however. The first is a defensive one: greaterautomation of software development and the emergence of other low-labor cost sources ofcompeting IT skills may lead to export growth falling or even reversing, as global demand forIndian programming services slows or falls due to automation. The second reason is that it maybe possible to do even better. Comparative advantage is not fixed, and countries can movetoward producing higher value-added goods and services as they grow, with favorableconsequences for long run growth. Applying endogenous growth models is not an automaticproposition, since results are sensitive to assumptions. For example, learning by doing inmanufacturing (including software production in this abstract conception) gives differentoutcomes than the assumption of a separate R&D sector that competes with manufacturing forskilled labor.

35 Note that, to the extent that India is providing intermediate goods or services in its software exports, the situationis more complex than that of standard trade theory, where only final goods are traded. Monopolistic competitionmodels such as those of Matsuyama are then relevant. For an attempt to calculate Indias possible comparativeadvantage, see Arora and Athreye (2002), Table 4.

15


17/32


If we accept the potential theoretical benefits of moving up the value added ladder, whatdoes this mean in practice for Indias software industry? One possibility is offering highervalue-added component services, involving design and strategy. Another is offering morecomplete packages or bundles of services. The latter differs from the former in that a highermanagement component is included in the package than in particular aspects of software

development, even if those require more technical skill. While the global economic slowdownthat began in 2001 will hamper these developments in India, there is no reason why Indiansoftware firms cannot enter such markets. The availability of a growing number of professionalswith combinations of engineering and management skills will help in this area.

36

Companies such as TCS have long-standing domestic and developing country consultingexpertise, but they may be less suited to compete in a crowded US market than Indias newsoftware giants such as Infosys and Wipro, or smaller newcomers.37 What is the possiblecompetitive advantage of these firms? Certainly lower costs will help, but these may be betterused to provide upgraded or broader services, rather than in competing on price alone. Inparticular, empirical research (Banerjee and Duflo, 2000) shows that reputation effects are quite

important for Indian software exporters.

38

Such companies may also develop a strong niche inother developing countries, where lower prices may be more important. A high tech slowdownin the US may actually aid the development of such firms in India, if it leads to some reversal ofthe brain drain.39 Otherwise, managerial and high-level technical skills may constrain movementup the value-added ladder. In some cases, including consulting as well as IT-enabled services,multinational firms have relaxed some of the managerial constraints through their own entry,importing managers as well as training local ones.

The Domestic MarketThe domestic market for IT products and services is not independent of the export

market. The nature of information goods in general is that they involve high fixed costs of

production and low marginal costs. While customization and service provision mitigate thisproperty, they do not negate it. Reputation and experience effects, on the other hand, enhance theimportance of economies of scale and scope. Hence it is important for Indian software firms tocompete simultaneously in domestic and export markets, in order to take advantage of theseeconomies. This is true even though the product-service mix that is being sold in differentmarkets is going to be somewhat different. Since Indian software firms can compete successfullyabroad, they should also be able to succeed in their own backyard.

40In fact, they have

advantages in the domestic market, knowing their customers better, and being closer to them.On the other hand, a poor domestic infrastructure, dependence on imported hardware, late mover

36

Similar considerations apply to IT-enabled services: see footnote 11.37 See Das (2001) for a discussion of the Indian e-business consulting market.38 Arora et al (2000) and Arora and Athreye (2002) discuss Indian firms efforts to signal quality by hiringengineers, and through international certification of their processes. They document the positive impact of the latteron value added per employee.39 For a description of how Infosys is struggling with the tech slowdown, to transform itselffrom a great code-writer to a one-stop technology services provider, see Bjorhus (2002).40 The reverse need not be true, as Arora and Athreye (2002) emphasize in their discussion. They also suggest thatthere are strong potential spillovers from the IT sector to other services industries, in the form of improvedmanagerial practices that have developed in IT and are easily applied to a range of services.

16


18/32


disadvantages, and lack of economies of scale and learning by doing, can all reduce or eliminateany advantage that Indian software firms might have over foreign competitors.

Two mitigating factors operate on potential disadvantages of Indian firms. First, some ofthe problems are faced by all firms, irrespective of location: for example, entering the market for

desktop operating systems in the face of Microsofts dominance is difficult, if not impossible, forany firm anywhere in the world. Second, the boundary lines between domestic and foreign canbe blurred when multinationals have Indian subsidiaries, particularly for IT or IT-enabledservices. In such cases, the effects on the local economy are not that different from when theseservices are provided by Indian firms. Two differences in the case of multinationals, however,are in profit repatriation and the creation of another brain drain channel, if Indian employees ofmultinationals can be assigned to other countries.

At the level of business software and software services, therefore, it seems that issues forthe domestic market boil down to the same concerns as for export markets. These are availabilityof the key inputs, namely various types of skilled IT personnel and managerial and marketing

skills. Location and ownership are not of direct importance, but are only proxies for whether theIT software and services provider has the right combination of people, knowledge, experienceand reputation to compete successfully. Liberalization in general means that all Indian firmsface the challenge of building such combinations of assets. The software industry happens to besignificant because it has developed independently of Indias traditional business houses, andhence mostly free of the bad habits those business groups developed over many decades ofoperating in noncompetitive environments.

We have focused the discussion so far on software. Hardware may offer additionalopportunities to Indian IT firms in the domestic market. In developed countries, theestablishment of the PC market took place before the Internet took off. In a good example ofcomplementarities, however, the growth of the Internet has increased the demand for PCs andother access devices. Internet access is probably the most attractive use for many potentialconsumers of IT in India but Internet penetration may not go far enough with hardware designedfor developed countries. Internet kiosks, with shared access, are a solution that has emerged forurban and rural areas, with the start-up cost for a basic kiosk having been brought down to under$2000. While Internet use is beginning to grow rapidly, the number of subscribers remainsminuscule, estimated at 1.8 million in December 2000. The main reasons for this backwardnesshave been the government long-standing monopoly, through VSNL, of the countrys Internetgateways, as well as the general poor state and high cost of the telecoms infrastructure. Theremoval of the VSNL monopoly in 2002 marks a process that began a few years earlier, withNASSCOM lobbying resulting in private ISPs being allowed to set up their own internationalgateways starting in 2000.

The possibility of designing and building lower-cost access hardware in India mayrepresent an opportunity for the domestic IT industry. While India has tried to develop adomestic hardware industry since the 1980s, it has not succeeded in establishing an industry thatis efficient and globally competitive.41 Much of the problem has lain with lack of scale andinfrastructure (as well as general restrictions on Indian business). For components such as

41 See Hanna (1994), for example, for further details.

17


19/32


sophisticated chips, this will continue to be the case, but, as noted earlier, Indian industry canbuild on existing capabilities in assembly of standardized components. If some of thesecomponents are designed specifically for the broader Indian market, for example to go into low-cost Internet and telecom access devices, as envisaged by the IIT Chennai group (footnote 27),where they are built may not be crucial.

42We note once more that Dell is a profitable company

because it serves targeted markets efficiently, not because it manufactures sophisticatedcomponents. Instead, management and infrastructure are the key inputs that are required.

Broad-Based GrowthAre a software industry that serves the domestic market as well as exporting, a hardware

industry that can produce low-cost access devices, and IT-enabled services for foreign markets,together enough for broad-based economic growth? The IT sector can directly contribute one ortwo percentage points to Indias growth rate, and this is not insignificant. The possible concernis that it will remain an enclave, exacerbating inequality and doing little for long-run growth.These concerns arise even in developed countries.

43

The argument for broad positive impacts is based on the kinds of models discussed at thebeginning of this section. To the extent that IT can have significant effects on the efficiency ofoperations in other industries, there are strong complementarities between the IT sector and therest of the economy. Examples of areas where increased efficiency may be possible includeaccounting, procurement, inventory management, and production operations.

44This is, of

course, the standard argument in the US for the virtues of the new economy based on IT. Toconnect our observations to the usual e-commerce jargon, we note that these benefits are situatedin the B2B arena. The difference for India is that it is starting from a much lower level of IT-adoption, and the potential gains may be higher. In fact, developing countries have theopportunity to leapfrog over older, more expensive approaches such as Electronic DataInterchange, which represent significant legacy investments in countries such as the US. This is

also a positive indicator for the future.

45

The argument of Kapur (2002), that Indias success insoftware exports has increased the confidence of Indians, may also be couched in terms of apositive shift in expectations helping to overcome a potential coordination failure.

A general concern with IT-adoption is job loss, and there is certainly the potential thatcertain kinds of clerical jobs will be eliminated or reduced in numbers. Unions in Indianindustries such as banking have opposed computerization for this reason. However, theevidence suggests that increases in other kinds of jobs as a result of IT use more than make upfor job loss, so that total employment is not a significant issue. In particular, IT-enabled servicespromise to directly generate employment much more significantly than activities such assoftware development. This leaves the issue of adjustment costs, and here severance pay rules or

42 Another example is the Simputer, designed by scientists from the Indian Institute of Science. The portable $200device will use some parts manufactured in Singapore, Linux-based software developed in India, and run on 3 AAAbatteries. See Khoo (2002).43 See, for example, Pohjola (1998) and Woodall (2000), as well as the references in footnote 3.44 These are all examples of what are also known as forward linkages, since IT adoption has positive impacts onthe operations of a range of industries. The effect of the growth of the IT sector on the provision of technicaleducation would be an example of a backward linkage. In either case, there is a complementarity at work.Evidence of some forward linkages in the informal IT sector is provided by Kumar (2000).45 See also Miller (2001) for a further discussion of the potential for B2B e-commerce in India.

18


20/32


government job adjustment assistance can be more effective and efficient than the current morassof detailed restrictions embodied in Indias labor laws. This is why Desai (2000) is right to stressthe importance of broader labor law reform if the benefits of growth in Indias IT sector are to befully realized.

In the context of complementarities, it is also important to recognize that these effects arenot just in terms of cost savings. IT implementation may enhance the quality of service beyondanything that is feasible through other methods. Furthermore, depending on who thecustomers are, the benefits may accrue to a broad cross-section of the population. Improvedefficiency in the stock market as a result of automated trading and settlement may benefit a smallsection of the population (though the indirect benefits of greater capital market efficiency may bebroader). The use of IT in rural banking and micro finance, however, can impact a much broadercross-section of the population. The evidence of pilot schemes such as the SKS InfoTech SmartCard project is encouraging. Handheld computers and smart cards can substantially reduce thecosts of making loans, as well as monitoring them. Reducing these transactions costs may turnout to be critical for the scalability and sustainability of micro finance schemes. This and similar

examples provide evidence against the enclave view of the IT sector in India.

Of course potential benefits do not necessarily translate into actual ones. Firms andmanagers can make mistakes in their IT investment decisions, but this is no different from anyother kind of investment. In a reasonably competitive industry, with sufficient informationavailable, there is always pressure to make the right decisions, rewards for those who do, andpunishments for those who do not. Indian industry must be allowed to follow this model torealize the potential benefits of IT. If it is discouraged from making such investments, thedomestic market for Indian IT will not grow, with negative consequences for the IT sector as awhole. This line of thought argues in favor of a sound competition policy, rather than anyspecific incentives, but we will explore this issue further in the next section.

We have focused mainly on the formal sector in assessing the impact of IT. Even if thegrowth of the IT has positive spillovers for other industries, this leaves out a substantial portionof the economy. We postpone a discussion of government use of IT to a separate section, butnow turn to issues of truly broad-based impacts of IT. There are several possible areas ofimpact. First, information processing may enhance efficiency in agriculture as well as inmanufacturing. While individual farmers cannot make IT investments, agricultural cooperativescan provide the institutional framework that allows farmers to benefit. For example, Chakravarty(2000) gives the example of IT use at milk collection centers in cooperative dairies. This permitsfaster and safer testing, better quality control, quicker and more accurate payments to farmers,and time savings for farmers in their deliveries. The falling cost of information processingmeans that such success stories can potentially be widely replicated.

The second impact is in the communication of information. Here the case studies arelegion. Farmers and fishermen can receive weather forecasts, market price quotes, advice onfarming practices, and specific training. Offers to buy or sell livestock, or other two-waycommunications are also possible. Some of this information dissemination and exchange is bestdone through voice media such as fixed or mobile telephones, while other types require the

19


21/32


capabilities of the Internet.46 Some evidence suggests, not surprisingly, that richer farmers andfishermen, as well as middlemen, are faster adopters of such technologies (The Economist,2001a), but falling access costs, through innovations such as those of the IIT Chennai group,should broaden information access and its benefits. Broad-based benefits of IT require broad-based access to the network, as we have discussed earlier.

A third area of impact is closely related to the second, but involves communication ofinformation in a more fundamental way. It is possible that IT-based delivery mechanisms canovercome traditional barriers to widespread delivery of education at all levels. We have noted theimportance of IT training itself. However, even basic education may be enhanced by the use ofIT. While it may seem paradoxical that delivery of basic education should rely on high tech,there is nothing new in this. The radio and television have been very successful distance-education media in the past, and computers and the Internet offer several advantages, in terms ofthe potential for interactivity, customization and sheer volume of material. Given the poor stateof basic education,

47while improved incentives for teachers and school administrators (either in

the public or private sector) will help, technology can play an important complementary and

even substitutive role. For example, TARAhaat (a semi-commercial subsidiary of an NGO), inattempting to develop a network of rural Internet centers in a district in Punjab, found that evenin the absence of reliable connectivity that would allow access to a variety of Internet-basedservices, it was able to tap into an underserved market for education in the vernacular medium inthe basics of computers and the English language.

48

The TARAhaat example illustrates several general points. First, in all attempts tointroduce IT to rural India in a manner that promotes development, sustainability is a key issue.The TARAhaat franchisee model offers important promise in this regard with respect toincentives and scalability, though there have been difficulties in implementation. Second, theexperiment validates the idea that IT costs have come down sufficiently to make rural IT servicesfinancially viable. Third, there is the issue of complementarities, both technological andpecuniary. One major roadblock for TARAhaat has been the poor quality of existing telecomsinfrastructure. This has severely limited the scope of services that its franchisees could offer,

49

and is an example of government failure. On the other hand, the provision of complementaryinputs such as financing and physical infrastructure, through subsidized loans from nationalizedbanks and the use of local government buildings, have been important in reducing startup as wellas operating costs. The most important complementarity emerged when the Punjab TechnicalUniversity quickly piggybacked on TARAhaats efforts, enhancing the franchisees initialfinancial viability through its own offerings of college-level IT education. This example suggests

46 Eggleston et al (2002) provide some quantitative evidence for the market efficiency effects of improved

communications and information transfer, using data from rural China.47 See Dreze and Gazdar, (1997) and the PROBE report (1999). One can also make a case for access to IT based onbroader notions of development, such as Sen (1999). That the poorest of the poor can benefit is borne out byinstances such as the famous in the hole-in-the-wall-computer (http://www.niitholeinthewall.com/home.htm).48 See Kaushik and Singh (2002) for more detail on the TARAhaat effort.49 An example from field research in Bathinda district of Punjab in December 2001 illustrates: a farmer told us hehad taken computer lessons at the TARAhaat kiosk, bought a home computer, and signed an Internet servicecontract so that he could exchange email with his brother in Toronto, Canada, as well as look for information onagricultural practices. All three IT-related products and services depended on basic telecom availability. See alsoPrahalad and Hart (2002).

20


22/32


how the kinds of coordination failure identified in the work of Matsuyama and others may beovercome.

GovernanceOne area where government can provide indirect support for the IT sector is by boosting

the domestic market though its own purchases. Initially, this seemed to follow the pattern of theintroduction of PCs in the US, with purchases of sophisticated equipment and software that satunderutilized on desktops. However, falling costs, increased understanding of what IT can do,and greater domestic expertise are starting to take the role of IT in government beyond thisbeginning.

There are two broad uses of IT for improved government functioning. First, back-officeprocedures can be made more efficient, so that internal record keeping, flows of information, andtracking of decisions and performance can be improved. Second, when some basic information isstored in digital form, it provides the opportunity for easier access to that information by citizens.The simplest examples are e-mailing requests or complaints, checking regulations on a web

page, or printing out forms from the web so that a trip to pick up the forms from a physical officecan be avoided. More complicated possibilities are checking actual records, such as landownership or transactions. Still more complicated are cases where information is submittedelectronically by the citizen, for government action or response.

The numerous examples of successful pilot e-governance programs include: Computer-aided registration of land deeds and stamp duties in Andhra Pradesh, reducing

reliance on brokers and possibilities for corruption Computerization of rural local government offices in Andhra Pradesh for delivery of

statutory certificates of identity and landholdings, substantially reducing delays50

Computerized checkpoints for local entry taxes in Gujarat, with data automatically sent to a

central database, reducing opportunities for local corruption Consolidated bill payment sites in Kerala, allowing citizens to pay bills under 17 different

categories in one place, from electricity to university fees E-mail requests for repairs to basic rural infrastructure such as hand pumps, reducing reliance

on erratic visits of government functionaries51

As in the broader case of using the Internet for communications and transactions,sustainability of e-governance initiatives is a significant issue. Since governments at all levels arefinancially strapped, the initial investments and ongoing expenditures for IT-based servicedelivery may act as a barrier to adoption as well as to long-run sustainability. However, ourinitial investigations suggest that the franchise model can be successful here. Low-cost ruralInternet kiosks, a tiered franchising model, and a suite of basic government access services forwhich users are willing to pay, are key components of what Drishtee, an outgrowth of theGyandoot project in Madhya Pradesh, is implementing in several parts of India.

52Cooperation

of local governments and subsidized financing have been important elements for Drishtee, as in

50 These two examples are from Bhatnagar and Schware (2000), which also provides broader examples, includingones driven more by the efforts of NGOs than governments.51 These three examples are from India Today (2000), which also lists several other similar projects.52 Further details of Drishtees efforts are in Kaushik and Singh (2002).

21


23/32


the case of TARAhaat, with the former being obviously critical in the case of Drishtee. It isimportant to note that once Internet access is available, its benefits are not restricted to e-governance. Individuals can obtain market information, training, job information, advice onfarming techniques, and so on, as discussed earlier in this section. This is certainly part ofDrishtees long run model.

Given the poor quality of governance in India, it seems that e-governance initiatives canprovide direct benefits to citizens, particularly those who are less well off (the rich in any casehire intermediaries to collect information, make payments, etc.). The preliminary evidencesuggests that the use of IT can increase transparency and accountability, simply by requiringinformation, such as basic complaints, to be logged completely and systematically. In thisrespect, the use of a non-governmental intermediary such as Drishtee may have advantages overpurely internal government initiatives, beyond that of financial viability.

5. Policy Environment

The overall goals of economic policy in India are standard: high growth together withmacroeconomic stability and poverty reduction. Balancing these goals is the difficult part. Forexample, incentives for exports, such as tax breaks, are designed to spur growth, but mayadversely affect the governments fiscal deficit. As quantitative controls have receded inimportance, such tax-subsidy policies have become more significant policy components. Thegrowth of Indias IT sector, and the success of the software industry in particular, has tended toskew policy toward the industry, with targeted incentives being implemented or recommended.

Targeting incentives to the software industry is not necessarily the best method topromote the industry, nor to achieve broader goals of growth and human development. Providingimplicit or explicit subsidies to the industry can introduce distortions, and it involves forgoing

other uses of funds, given the severe budget constraint that the government faces. Broaderpromotion of the IT sector also suffers from some of the same problems. Investing heavily ingovernment-sponsored IT-related training is problematic when basic education in India is sopoorly provided by the public sector. Policy goals for the IT sector might be better met byfocusing on infrastructure provision, enabling the private sector to play a role here as well. Thetelecoms sector is a case in point.

The historical case for regulation or nationalized provision in telecoms was based oneconomies of scale, implying that competition would be unstable or inefficient. Technologicalchange has removed this justification in significant portions of the telecoms value chain bylowering fixed costs and adding new technological options, allowing competition to become

feasible. Since monopoly may persist in portions of the value chain (i.e., portions of thenetwork), regulation of interconnection charges may still be required, to maintain a level playingfield. However, directly managing technology choices and competition is not easily justified oneconomic grounds. The broad policy goals of promoting competition and innovation in theprovision of telecoms infrastructure and services are moving in the right direction.

Broad-based growth of Indias IT sector will depend on improving the telecomsinfrastructure, and on training enough people for the sector and using them effectively and

22


24/32


efficiently. For telecoms, the regulatory framework is crucial, whereas for human resourcedevelopment and use, the labor laws matter greatly. It may also be noted that laws that directlyconstrain manufacturing remain on the statute books, and adversely affect areas such asmanufacturing or assembling hardware the problem here is one that still affects Indianmanufacturing in general.

53

In India the regulatory institution for telecoms is the Telecom Regulatory Authority ofIndia (TRAI), which was constituted in 1997 and given greater and clearer authority in 2000.

54

The scope of the TRAI includes establishing quality of service parameters, monitoringcompliance, examining technology choices, and so on.55 It is supposed to establish a levelplaying field and encourage competition, but it has lacked clear authority precisely where itneeds it the most, in setting entry fees and some interconnection charges. Unfortunately,bringing quality of service, technology choice, and universal service obligations (USOs) into theregulatory mix only serve to muddy the waters, and divert attention from the central task ofenabling effective competition.

USOs are being built into licensing deals for private service providers. These take theform of quantitative targets for installing rural telephones, and funds created through a form oftax on basic service, to be used for proposed subsidies for rural users. It is not clear thatnumerical targets have any use at all, when licensing and interconnection fees make ituneconomical for local access providers with lower-cost technology to enter. The distinctionbetween rural users in general, and shared access through Internet kiosks is crucial, but has notbeen accepted by majority of a TRAI committee that reported on the USO.56 Our owninvestigation of the startup problems of TARAhaat and Drishtee suggests that a narrowlytargeted subsidy for enabling reliable telecoms access (including solutions such as that of the IITChennai group) to Internet kiosks makes most sense.

Desai (2000) examines the problems of labor laws, using the Report of the Subject Groupon Knowledge-Based Industries (2000) as his starting point. The report calls for exemption forthe IT sector from a broad set of rules relating to labor, including provisions relating to overtime,working conditions, restrictions on contract labor, and dismissal of workers. Interestingly, if ITworkers are in short supply, they should be able to negotiate terms that are attractive enough tomake the labor laws redundant. Desai suggests that the main function of the labor laws in thissector is to enable government labor inspectors to demand bribes. He also argues for broaderreform of labor laws, and rightly points out the potential for distortions if one sector is given anexemption. He also acknowledges the political difficulties of more comprehensive reform. Inthis case, the IT sector may usefully serve as the thin edge of the wedge that begins cutting downsome of the worst problems with Indias labor laws, in particular the lack of permitted flexibilityin contracting. The development of IT-enabled services will be a litmus test of the changing roleof labor laws in India.

53 I am grateful to P.D. Kaushik (personal communication) for this point by his count, there are over 400 centralgovernment statutes governing manufacturing, as well as numerous state laws. He also notes that the softwareindustry escaped these constraints partly by not being recognized by the government as an industry.54 See Dossani and Manikutty (2000) for details.55 See for example, the paper by M.S. Verma, Chairperson of the TRAI (Verma, 2000).56 See the report at http://www.trai.gov.in/recom.htm, and especially the two Annexes, which are a dissentingcomment by Rakesh Mohan, and the rest of the committees response. See also Dey (2000) and Singh (2002).

23


25/32


Our examination of the role of the IT sector in broad-based development also suggeststhe importance of the financial sector. While it seems that large IT firms can rely on retainedearnings or the stock and bond markets for growth, startups need a venture capital industry that isjust beginning to emerge. As in the case of labor laws, one can argue that the policy environment

must be geared toward industry in general, and not just the IT sector. Kapur and Ramamurti(2001), for example, note that industries such as biotech, chemicals, media and entertainment,and construction all require knowledge services that go beyond the basic definition of IT-enabledservices. In all these cases, venture capital may be a significant fuel for entrepreneurial energy.

Some of the greatest difficulties face small-scale entrepreneurs, who have been protectedby reservations, but who do not necessarily have easy access to the right kinds of help they needas startups. Again, our research on the experience of the local franchisees of TARAhaat andDrishtee suggests that the nationalized banking sector, with its system of directed credit, andsimultaneous forced holdings of government and PSU loans, which have left bank portfolios inbad shape, is not well placed to provide small-scale financing of this kind. In particular, in the

case of TARAhaat, difficulties in obtaining startup capital from banks appeared to be a majorimpediment to rapid expansion of the franchising scheme, even within a small geographic area.57

This is where the overall macroeconomic problem of the fiscal deficit appears to trickle down allthe way to the village, with a negative impact on development.

Our general assessment of the policy environment is therefore that, with one exception,the policy problems are mostly general ones, and not specific to IT. Problems in financial marketinstitutions, labor laws, and regulation in general are best dealt with from an economy wideperspective, as we elaborate on in the conclusion. The exception is the case of rural telecoms andInternet access, where it seems that narrow targeting of limited subsidies (through waivingcertain fees rather than explicit payments) for startup costs may be worthwhile in generatinggrowth of communications, and of enterprises that use communications.

6. Conclusion

To conclude, we will briefly consider general microeconomic and macroeconomic policyissues, and implications for the IT sector. The central areas of Indias policy reforms have beenreplacing quantitative trade restrictions with tariffs, lowering effective levels of protection,removing an area of discretionary controls on private sector investment, and creation of modernfinancial markets. Standard examples of where these reforms can be built upon, to furtherstimulate growth, include removal or relaxation of obsolete small-scale sector reservations andsize restrictions, privatization of inefficient state-owned enterprises, rationalization of tax-

subsidy policies and tax administration, and relaxation of severe labor market restrictions. Thislist can be characterized by its emphasis on improving the efficiency of the mechanisms withwhich the government directly affects the private sector. The entire Indian economy, not just theIT sector, can presumably benefit from such reforms, which will reduce distortions of privatesector behavior.

57 Drishtee was able to avoid this problem to some extent, with smaller-scale kiosks that allowed poorerentrepreneurs to avail of targeted government loan schemes.

24


26/32


A second area where attention is required may be characterized as enabling reforms.These include reforms of contract law and judicial institutions; financial sector regulatoryinstitutions; telecom sector regulatory institutions; infrastructure such as electric power, roadsand ports; and systems of education and training in general. Again, the benefits of such reformsare potentially quite general, and not restricted to any one sector of the economy.

A third area of policy is macroeconomic management. While Indias record here is quitegood, it needs to make a transition in its policy institutions here as well, since removing detailedmicroeconomic controls requires changes in the regulatory modes of macroeconomicmanagement. Perhaps the area that has received the most attention is policies towardinternational capital flows and their implications for exchange rate management. Desai (2000)has suggested that large projected increases in software exports could create a Dutch diseasephenomenon,58 in which a resulting exchange rate appreciation hurts other sectors, and revenuesfrom exports are wastefully spent. Several factors mitigate this concern: the likelihood thatexport revenue growth will slow down; the potential linkages that exist between software, the ITsector as a whole, and the broader economy (unlike natural resource extraction enclaves); and a

better understanding of exchange rate management than existed 25 years ago, when thephenomenon first was identified and labeled. Thus, while exchange rate policy is certainlyimportant in general, the growth of the IT sector will not necessarily raise special concerns.

Given that there is plenty that remains to be done in terms of overall economic policyreform, are there areas where the IT sector deserves special attention? The answer we havegiven in this paper, with one exception, is no. Special subsidies or export incentives are likelyto be inefficient ways of stimulating the growth of the IT sector, or of positive spillovers for therest of the economy. Similarly, special central government initiatives to increase the availabilityof IT training and related education are also likely to represent a mistargeting of scarcegovernment resources. The same stricture applies, to some extent, to state government policies toencourage the IT sector.59 The government may be better off removing general restrictions todoing business, as well as providing an enabling institutional infrastructure (appropriate laws andregulations), rather than attempting to target the IT sector through a form of industrial policy.

The exception lies in the telecom sector, which has particularly strong complementaritieswith the broader IT sector. Policies to achieve development goals would do better to emphasizeremoving barriers to innovations that will support lower-cost access to telecom networks of allkinds (wireless and fixed, voice and data). Very specific, targeted, startup subsidies to enablewidespread, shared access to telecoms and Internet in rural areas are likely to have high socialreturns, since it appears that financially sustainable franchise models exist. These high returnsinclude better governance, as well as knowledge that is an important input into empowerment,or development as freedom (Sen, 1999). In this respect, we would argue that rural IT access is

58 An excellent explanation of Dutch disease is by John McLaren, at www.columbia.edu/~jem18/teaching/pepm/dutchdis.pdf. McLaren clarifies the source of concerns that are associated with Dutch disease, includingexacerbation ofpriordistortions, and of inequality.59 Bangalore in Karnataka is well known as a regional IT center in India, having developed initially without muchexplicit government support. The governments of Andhra Pradesh (Eischen, 2000) and Tamil Nadu (Bajpai &Radjou, 1999, and Bajpai & Dokeniya, 1999) have led in attempts to establish IT-based industries with consciousgovernment policies. Other state governments, such as Punjab (see www.dqindia.com/mar1599/news.htm) arefollowing suit, with mixed success.

25


27/32


an important complement to and enabler of local government reform in India (Rao and Singh,2000).

Our goal in this paper has been to assess the possible role of Indias IT industry as adriver of higher economic growth in India, without exacerbation of inequalities or creation of

instability. Our conclusion is cautiously positive. While projections for software exports may beover optimistic, complementarities or spillovers in the domestic market, including increasedgovernment and business use of IT, are likely to be strong. For this rosy scenario to play out,however, continued broad economic reforms will be important, as well as reforms in the telecomsector that promote competition and innovation in providing last-mile access.

References

Aggarwal, Balaka B. (2001), Faculty scarcity at IITs threatens knowledge capital, March 19,http://www.ciol.com/content/news/trends/10103902.asp.

Arora, Ashish and V. S. Arunachalam, Jai Asundi and Ronald Fernandes (2000), Theglobalization of software: the case of the Indian software industry. A report submitted to theSloan Foundation. Carnegie Mellon University, Pittsburgh PA.http://www.heinz.cmu.edu/project/india/publications.html.

Arora, Ashish, V. S. Arunachalam, Jai Asundi and Ronald Fernandes (2001a), The IndianSoftware Service Industry,Research Policy, 30, 1267-1287.

Arora, Ashish and Suma Athreye (2002), The Software Industry and Indias EconomicDevelopment,Information Economics and Policy, 14, 253-273.

Arora, Ashish, Alfonso Gambardella and Salvatore Torrisi (2001b), In the footsteps of theSilicon Valley? Indian and Irish Software in the International Division of Labo

Documents

IT India Cornell Rev